Fractured reservoirs contain a large fraction of the world supply of oil. For viscous crudes, steam is the most successful technique and field tests indicate that steam has the best potential to recover significant amounts of oil from fractured reservoirs. Unfortunately, there has been little laboratory work done on steam injection in such systems.
The experimental system discussed here was designed to understand the mechanisms involved in the transfer of fluids between the matrix rock and the fracture as a result of steam injection. Both continuous and cyclic steam injection experiments were performed on a fractured laboratory system. Saturations were measured in-situ both in the fracture and the consolidated matrix by a CT scanner. The results indicated that there was no steam saturation in the matrix, and that conduction was the dominant heat transfer mechanism.
Numerical simulations were used to model both continuous and cyclic steam injection experiments. To model heat losses, heat loss models in the simulator had to be adjusted based on the analysis of the heat losses from the laboratory system with analytical models. After this adjustment, the results from the simulations agreed well with the experiments.
When pressure cycling was applied in the simulations with no external heat losses, a considerable amount of steam saturation was observed in the matrix.
While the experiments were done with water and steam, simulation runs were also performed for the laboratory system with oil present. Again, steam only flowed in the fracture. Oil recovery was found to be mainly caused by water imbibition into the matrix and heat conduction.
Results of this work should be useful in modeling matrix/fracture transfer in dual porosity thermal models.
Fractured reservoirs are estimated to contain 25-30 % of the world supply of oil. Steam injection is required for most of the reservoirs containing heavy oils and tars. There have been quite a few field studies on steam injection for fractured systems (Sahuquet and Ferrier (1982), Britton et al. (1982), Stang and Soni (1987), Closmann and Smith (1983), Duerksen et al. (1984), Couderc et al. (1990) and Hartemink et al. (1995)). Most of these applications were successful or promising.
Thomas (1964), Lesser et al. (1966), Abdus Satter (1967) and Satman (1988) presented theoretical models for conduction heating of formations by injecting a fluid through a high permeability streak or fracture. Van Wunnik and Wit (1992) developed a detailed analytical model to study the improvement of gravity drainage by steam injection in a fractured reservoir containing heavy oil. Pooladi-Darvish et al. (1994) developed analytical solutions for heat flow and nonisothermal gravity drainage from a block surrounded by fractures filled with steam. In addition to these analytical models, there have been some numerical simulation studies for steam injection in fractured systems (Pruess and Narasimhan (1985), Lee and Tan (1987), Chen et al. (1987), Pruess and Wu (1989), Nolan et al. (1980), Abad and Hensley (1984), Lin (1988), Briggs (1989), Jensen et al. (1992) and Oballa et al. (1993)).
The Wilmington Steamflood of Union Pacific Resources Co. (UPRC) at Long Beach, CA was initiated in 1989, in a previously waterflooded reservoir. Average initial reservoir oil saturation, at the start of the steamflood was 35%.
Field production data were studied, to derive an overall energy balance for the steamflood to calculate the steamflood capture efficiency and predict further steamflood performance. Heat-losses due to produced fluids were calculated. Predicted production schedules from the model were history-matched with field production data.
All steamflood calculations were carried out using a PC-based spreadsheet program. The major results were as follows:
- The capture efficiency of the Wilmington steamflood was calculated at 60%. This is an acceptable value, taking into account the fact that the reservoir had previously been waterflooded to a low oil saturation of 35%. - The calculated heat balance showed a high heat-loss, not only to adjacent formations, but also through produced fluids. Of the cumulative heat injected up to the time of the study, 21% had been lost to vertical conduction and 21% through produced fluids.
- Predicted production schedules indicated that up to 43% of the oil in place (at steamflood initiation) would be recovered by the steamflood.
The Wilmington Oil field, Los Angeles County, California is the third largest in the U.S., after Prudhoe Bay and the East Texas fields, on the basis of cumulative oil production, with a total of 2.4 billion barrels produced. A good description of the initial pilot steam flood carried out by UPRC in this field from 1982-1989 is given by Lim et al., 1993. The pilot 20 acre steamflood is now part of the main 130 acre Wilmington Steamflood. The portion of the field studied had initially been waterflooded to a low oil saturation of 35% before the main steam flood was initiated in 1989.
The steam flooded reservoir is unconsolidated sandstone in the Tar Zone (T, D1 and D3 members) with average gross and net thicknesses of 170 feet and 128 feet respectively. Average reservoir pressure was estimated at 350 psig in 1993. The steam flood project has a surface pattern area of 129.73 acres divided into approximately 17 individual seven-spot patterns, each of 7.5 acres. Average well depth is 2500 feet. The oil in place at steam flood initiation was calculated at 18 million barrels with a movable oil volume of 12.9 million barrels, assuming an initial oil saturation, Soi, of 35% (at steam flood initiation), a final oil saturation Sor of 10% and an average porosity of 40%.
To simulate steam injection in fractured reservoirs either double porosity or double permeability models are used. Both incorporate a matrix-fracture transfer term in the mass and energy balances. Due to the difficulty in modeling the physical processes taking place in the matrix-fracture transfer and the lack of experimental data, the matrix-fracture transfer term is not fully understood. This is especially true for nonisothermal processes.
Fine grid simulation results are used to design a 3-D laboratory matrix-fracture model to study the matrix-fracture transfer function for steam injection. A CT-Scanner will be used to measure the three-phase insitu saturations in the fractured model. The flow parameters were determined by using the results of several simulation runs. Among these are the steam injection rate, maximum expected pressure in the system and the number and the locations of the injection and production wells needed to clean and saturate the model for each run. Analytical heat transfer models were used to determine the heat losses from the model.
Fine grid simulations were used to investigate the sensitivities of the flow process to matrix-fracture properties such as capillary pressure and relative permeability. The simulations showed that matrix water-oil capillary pressure increased the oil recovery by increasing the water imbibition rate. On the other hand, fracture water-oil capillary pressure decreased the water imbibition rate and recovery. Matrix gas-oil capillary pressure had no effect on the oil recovery but fracture gas-oil capillary pressure had a positive effect on recovery by allowing flow of steam into the matrix thus increasing oil mobility. The effect of matrix relative permeability was found to be less important.
The experimental design was modified based on the numerical results. We give details of the experimental apparatus and show some results from preliminary experiments.
Several flow models have been developed in the literature for isothermal processes in fractured media. The models used in thermal simulators are generally extensions of models for isothermal processes. They are classified into two groups, dual porosity and dual permeability models.
All dual porosity models assume that fractures constitute the main path for fluid flow, and matrix blocks act as sources or sinks to the fracture network. In the basic dual porosity model, matrix and fracture communicate through a single exchange term in the flow equations. Kazemi et al. (1976) developed a three-dimensional 2-phase model to simulate such a fractured system. The simulator equations are two-phase extensions of the single-phase equations derived by Warren and Root (1963) Thomas et al. (1983) extended the model to 3-phase flow in fractured systems.
In the multiple interacting continua model which is a different type of dual porosity model, the matrix is divided into nested volume elements which communicate with each other. In this model, flow of fluid and heat between matrix and fracture is transient. The model was first used in geothermal reservoir simulation by Pruess and Narasimhan (1985). Oilman (1986) applied this model to hydrocarbon reservoirs. Pruess and Wu (1989) developed a semianalytical model to simulate matrix-fracture transient flow.
Dutra and Aziz, (1992) developed an analytical transfer function, which results in a model having the same form as single-porosity models.
SPE Members *Now with Union Pacific Resources Co.
In steam drives and certain gas injection projects, the use of foamers to reduce steam and gas mobility has become a common way of improving sweep efficiency. Although the process is widely used, no effective means have been available to optimize it because there is no convenient flow model that properly describes the displacement characteristics of foam flow the way Buckley-Leverett theory describes simple two phase flow. To attempt to develop a useful model for foam displacement, a number of experiments were run using surfactant laden water displaced by air at a constant rate in linear sandpacks. To decouple the displacement problem from the heat transfer mechanisms, no steam was used in these experiments. The saturations were measured in situ using the Cat Scanner; and, in addition, production and pressure histories were measured for the overall system. First, water containing various concentrations of surfactant was displaced by air. As expected, these displacements follow Buckley-Leverett theory with saturation distributions growing linearly with volume injected. Also, as expected, the effective mobility of the air was a strong function of surfactant concentration as well as the air/water saturations. In addition, irreducible water saturations were found to be a strong function of surfactant concentration. Analytic expressions were developed to describe the relative permeability relationships as a function of saturations and surfactant concentration. Next, experiments were run where surfactant laden water and air were injected simultaneously into a porous medium filled with pure water. This process more closely resembles field applications. The displacement fronts were spread more than when the water in place contained surfactant. When multiple fronts of different concentrations were used to match the displacements, the calculations predicted the greater spread observed experimentally, but did not properly ape the effect of distance moved. These calculations, for example, predicted that the saturation dissipation length would double as twice as much fluid was injected. The experiments showed less growth than this. This slower growth rate could be the result of a dispersion term in the displacement. An equation was developed which included such a term and its predictions were found to match the experiments. The in situ saturation distributions were matched with distance and time as well as the overall recovery histories of the displacements. Under the conditions of these experiments, about 20% of the total spread of the displacements was caused by dispersion, and the rest was caused by Buckle-Leverett spreading. These modeling efforts have already resulted in insight for practical application. They predict that a large slug of surfactant solution at low concentration injected ahead of the steam will have a greater mobility control effect than will the same amount of surfactant injected simultaneously with steam.
There are various types of gas drives used for oil and gas recovery front petroleum reservoirs. Some are simple injections of gas such as methane; some are miscible gas injections such as hydrocarbon or CO2 miscible flooding; and some are steam injections. All these injection methods tend to override due to gravity, and channel and finger due to viscous instability and heterogeneity of the porous medium. Anything that would help alleviate these problems would improve the ultimate recovery of gas and steam drives. In many laboratory and field experiments, foam has been used to compensate for these effects.
In-situ combustion is the most energy efficient of the thermal recovery methods. In light oil reservoirs, too little fuel may be deposited thus making combustion impossible while in heavy oil reservoirs too much fuel may be deposited thus ruining the economics of the process. A research program has been initiated to try to solve these problems. Water soluble metallic additives ere tested to attempt to modify the fuel deposition reactions.
In a previous paper, results were reported from kinetics experiments run on Huntington Beach, California and Hamaca, Venezuela crude oils in the presence of aqueous solutions of metallic salts. While the presence of copper, nickel and cadmium had little or no effect; iron, tin zinc and aluminium increased fuel laydown for Huntington Beach oil. The results were similar for the heavier Hamaca oil.
This paper describes thirteen combustion tube runs using four different crudes. In addition to the above two crude oils, a 12 degree API and a 34 degree API Californian oil were tested. The metallic additives iron, tin and zinc improved the combustion efficiency in all cases. Fluctuations in the produced gases were observed in all control runs but disappeared with the iron and tin additives. The front velocities were increased by the metallic additives. Changes were also observed in H/C ratio of the fuel heat of combustion, air requirements and density of the crude produced. The amount of fuel deposited varied between the produced. The amount of fuel deposited varied between the oils. For Huntington Beach oil, the amount of fuel increased in the order: zinc, control, tin and iron while for the Hamaca crude the order was: control, iron and tin. The most interesting result occured with the light California oil. The control run showed that we were unable to propagate a combustion front while with iron additive a good combustion was achieved.
To date we have not been able to find a suitable additive to reduce fuel deposition. Iron and tin salts seem suitable agents to increase fuel when that is needed.
In-situ combustion is after steam injection, the most widely used thermal recovery technique. In this process, air or oxygen is injected and burns part of the oil which is used to generate a burning front which propagates in the reservoir. Oil production is improved by the hot gases generated during the burn. production is improved by the hot gases generated during the burn. Although the main application of thermal recovery is to re-t cover viscous heavy oils, a broad range of oils and reservoirs are potential targets for these recovery methods. The major constraint limiting the applicability of in-situ combustion is the amount of fuel formed on the reservoir matrix ahead of the combustion zone. If insufficient fuel is deposited, as can be the case for light oils, the combustion front will not be self sustaining and will die from lack of fuel, Conversely, if excess fuel is laid down, the front advance will be slowed and the quantity of oxidizing gases required to sustain combustion will be uneconomical. The amount of fuel formed and the velocity of the combustion front are governed by the kinetics of oxidation and pyrolysis reactions of the crude oil in the porous matrix Catalytic compounds affect the kinetics of the reactions and so can influence the amount of fuel formed. If the proper catalyst can be introduced in the reservoir to modify the oil's tendency to deposit fuel, then in-situ combustion could be made feasible for a broader range of crude oils and reservoirs. Such an application was first presented in 1985 by Racz. Recent work at Stanford University (Shallcross et al. and Baena et al.) showed that water soluble metallic additives can change the kinetics of the combustion reactions.
A general method has been developed to compute steam distillation yield and to quantify oil quality changes during steam injection. It was found that steam distillation data from the literature can be correlated with the steam distillation yield obtained from the DOE crude oil assays. The common basis for comparison was the equivalent normal boiling point. Blending of distilled components with the initial oil results in API gravity changes similar to those observed in several laboratory and field operations.
In steam injection the principal mechanisms responsible for additional oil recovery are thermal expansion of oil, visosity reduction and steam distillation. Steam distillation yields can be significant, even for heavy crudes. In reservoirs containing light oil, steam distillation is the major mechanism contributing towards improved recovery.
Steam distillation is the main mechanism which reduces the residual oil saturation behind the hot water front during a steamflood. At a given steam injection pressure or temperature, the residual oil saturation is essentially composition dependent. An increase of 2 degrees to 4 degrees API in the gravity of produced oil has been observed both during laboratory produced oil has been observed both during laboratory experiments and field tests. The residual oil in the steam zone has a high content of heavy components. Similar observations have been made during in-situ combustion field tests.
A major objective of this study is to investigate changes in quality of produced oil compared to the initial oil. Existing laboratory and field data will be interpreted in the light of the findings.
Quantification for oil quality changes during steam injection requires understanding of the major phenomena associated with the process. Steam distills lighter components, leaving behind heavier residuum. Blending of the lighter distilled components with the initial oil results in produced oil with higher API gravity. The main objective is to quantify these phenomena. For a given steam distillation condition, questions which phenomena. For a given steam distillation condition, questions which should be answered are:
* What is the steam distillation yield? * What is the quality or API gravity of the distillate? * How much heavier does the residuum become? * What API gravity change can be expected when the distilled components are mixed with the initial oil?
The Department of Energy (DOE) Crude Oil Analysis Data Bank will serve as the main source of information. A basis of correlating the DOE data with steam distillation data will be established. The results will be verified by comparison with experimental data published in the literature.
The DOE Crude Oil Analysis Data Bank contains crude oil assays for all major reservoirs in the United States. These crude oil analyses were obtained by the Bureau of Mines routine method distillation. Table 1 shows the distillation cuts and the calculated equivalent normal boiling points (ENBP).
Data from 454 California crude oil samples from the DOE Crude Oil Data Bank were analyzed. Fig. 1 shows the cumulative distillation yield as a function of API gravity for Cut Number 12. The distillation temperature is 437 degrees F at a pressure of 40 mmHg. The ENBP is 635 degrees F. A linear relation was pressure of 40 mmHg. The ENBP is 635 degrees F. A linear relation was found to fit the data:
The purpose of this study is to develop exact analytic solutions for the pressure response of a finite conductivity fracture. This model should be able to verify the existence of various flow regimes found in earlier studies. It is hoped that this solution could be modified to give simplified expressions for well pressures for all times and all fracture conductivity ranges.
The present work poses and solves the problem of a vertical finite conductivity fracture of elliptical cross-section. The flow within the fracture is assumed to be incompressible and the reservoir is assumed to be infinite. The elliptical fracture geometry was chosen to facilitate the expression of fracture and reservoir pressures as eigenfunction expansions.
The solution is obtained by expressing the reservoir presure as a series of Mathieu functions, and the fracture pressure presure as a series of Mathieu functions, and the fracture pressure as a series of cosines. The coefficients in these series satisfy an infinite set of linear relations, termed Fredholm sum equations. Exact solutions to these sum equations are obtained in forms which resemble continued fractions of summations, or equivalently, which require iteration of rational forms. A great deal of effort has been expended to speed the calculation of the solutions, however, only partial success has been achieved.
The solutions become increasingly difficult to compute as time decreases. So, approximate solutions for well pressures are given for extremely low values of time. These solutions indicate that behavior of an elliptical fracture is essentially the same as that of a rectangular fracture. Indeed, the well pressures calculated in this work are quite dose to those for a rectangular fracture. Generally applicable simplified well solutions have not been found.
The topic of the pressure response of fractured wells is not new. Many models have been investigated which consider various aspects of the problem. However, these models either consider only a part of the problem, or they allow only approximate solution of the governing equations.
The most comprehensive model that has been investigated is the rectangular finite conductivity fracture model developed by Cinco and coauthors in a number of papers. The two most important of these are Cinco et al., where the model is proposed and the governing equations solved, and Cinco and proposed and the governing equations solved, and Cinco and Samaniego, where the behavior of the solution is investigated.
The solution procedure used in the first of these papers was a numerical solution of an integral equation. The method is computationally intensive, and, since it is numerical, yields only approximate results. The pressures computed using this method appear to be accurate, although it is difficult to say just how accurate they are.
In Cinco and Samaniego a simplified model was analyzed and used to identify the bilinear flow regime. The regimes of fracture linear flow and reservoir linear flow were also examined. The fracture linear flow regime results from expansion of fluid in the fracture. This regime is of too short a duration to be of practical use and so little is lost by assuming the fracture flow to be incompressible.
The intent of the present work is to present a model which depicts the flow of fluids into and through a finite conductivity vertical fracture. We seek an exact solution which can be used to find pressures anywhere in the fracture/reservoir system.
The CT imaging technique together with temperature and pressure measurements were used to follow the steam pressure measurements were used to follow the steam propagation during steam and steam foam injection experiments propagation during steam and steam foam injection experiments in a three dimensional laboratory steam injection model. During the design period, the advantages and disadvantages of different geometries were examined to find out which could best represent radial and gravity override flows and also fit the dimensions of the scanning field of the CT scanner. As a result of this analysis a 3D rectangular box with dimensions 20 x 20 x 7.5 cm was constructed. This box simulates one quarter of a five spot pattern. Aluminum, Teflon and Fiberfrax were chosen as supporting and insulating materials. Teflon was placed between the porous medium and the aluminum shell so that the rate of heat transfer in the porous medium would be much faster than that in the aluminum during a steam injection run. During experiments, steam was injected continuously at a constant rate into the water saturated model and CT scans were taken at six different cross sections of the model. Pressure and temperature data were collected with time at three different levels in the model. CT pictures and three dimensional temperature distributions were compared and analyzed in terms of observed steam zone at each section. To do that, CT numbers within the scan section were used to determine the steam and water zones, and with the aid of x-ray pictures the position and propagation of the steam zone were determined. In addition, using the three dimensional temperature distribution measurements at the same times, steam displacement fronts could be drawn at the scan section locations. These pictures and drawings were used to compare the results obtained from classical temperature-pressure monitoring and from CT scans.
During steam injection experiments the saturations obtained by CT matched well with the temperature data. That is, the steam override as observed by temperature data was also clearly seen on the CT pictures.
During the runs where foam was present, the saturation distributions obtained from CT pictures showed a piston like displacement. However, the temperature distributions were different depending on the type of steam foam process used. During the experiment which included non-condensible gas (nitrogen) injection, the temperature distributions, contrary to the saturation distributions, still indicated the presence of steam override, although the override was reduced by the foam. However, when there was no nitrogen the temperature distributions followed the saturation distributions. This may possibly indicate that the nitrogen foam ahead of steam foam caused the difference between temperature and saturation distributions. These results clearly show that the pressure/ temperature data alone are not sufficient to study steam foam in the presence of noncondensible gas.
CT scanners, although they were invented for medical purposes in 1972 have been heavily used in petroleum engineering applications. A CT scanner in a petroleum research laboratory can be used as a useful tool for insitu saturation and porosity determinations and investigations of coreflood experiments, comparison of simulation results with coreflood experimental and EOR applications. Although CT scanners have been used to investigate the mechanisms of some EOR processes, these experiments were mainly carried out in linear models, with only a few in 3D ones. This leads to poor representation of phenomena such as gravity override and channelling.
Shallcross, D.C. (Stanford U. Petroleum Research Inst.) | de los Rios, C.F. (Stanford U. Petroleum Research Inst.) | Castanier, L.M. (Stanford U. Petroleum Research Inst.) | Brigham, W.E. (Stanford U. Petroleum Research Inst.)
Experiments were performed to study the effects of various additives on theoxidation kinetics of Californian and Venezuelan oils. Aqueous solutions of 10metallic salts were indexed with sand and Huntington Beach, CA, oil. Themixtures were subjected to a constant flow of air and a linear heating schedulewhile the effluent gases were analyzed for composition. The variation in theoxygen consumption was analyzed with a model of three competing oxidation reactions. Values for the important kinetic parameters for the three reactionswere obtained for each additive. Iron and tin salts were found to enhance fuelformation, while copper, nickel, and cadmium salts bad no significant effects.Other experiments with a heavy Venezuelan oil showed that, contrary to earliersuggestions, the use of a ketal did not decrease fuel formation.
The major constraint limiting the applicability of the in-situ combustionoil recovery process is the propensity of the reservoir oil to form fuel in thereservoir matrix ahead of the combustion zone. If insufficient fuel isdeposited, as is often the case for light oils, the combustion front will notbe self-sustaining and will die out quickly. Conversely, if excessive fuel isdeposited, the advance rate of the combustion front will be slow and thequantity of the oxidizing gas (usually air) required to sustain combustion willbe uneconomically high. If methods can be developed to modify the tendency foroil to deposit fuel, then the in-situ combustion process could be made feasiblefor a wider range of crude oils. The amount of fuel formed and the velocity ofthe combustion front are governed by the kinetics of the oil oxidation andpyrolysis reactions. Catalytic and organic compounds affect the kinetics ofthese reactions and so influence the amount of fuel formed. Oil oxidationduring in-situ combustion involves numerous competing reactions occurring overdifferent temperature ranges. A number of experimental studies haveinvestigated the kinetics of these oxidation reactions. A common procedureinvolves subjecting a mixture of oil, water, and sand to a linear heatingschedule. The effluent gas produced by passing air through the mixture duringthe heating process is analyzed for its oxygen and carbon oxides content. Thesestudies established that the overall oxidation mechanism of crude oils inporous media may be represented by grouping the reactions into three classes ofcompeting reactions occurring over different temperature ranges: (1)low-temperature oxidation (LTO) reactions, which are heterogeneous (gas/liquid)and produce no carbon oxides; (2) medium-temperature oxidation produce nocarbon oxides; (2) medium-temperature oxidation (MTO), fuel-formationreactions, which are homogeneous (gas phase) and involve the oxidation of theproducts of distillation phase) and involve the oxidation of the products ofdistillation and pyrolysis; and (3) high temperature oxidation (HTO),fuel-combustion reactions, which are heterogeneous, in which the oxygen reactswith the fuel formed during the medium-temperature reactions. LTO reactions canoccur in high-permeability streaks, which enable oxygen to travel so rapidlythrough the combustion zone that it contacts insufficient fuel for completeoxygen utilization to occur. LTO results in the production of partiallyoxygenated compounds such as aldehydes, ketones, and alcohols. This partialoxidation increases the viscosity and boiling range of the fluid. The regionahead of the combustion front is heated by conduction, by convection of thecombustion gases, and by condensation of volatiles and steam generated frominterstitial water. As the temperature rises, the crude oil undergoes threeoverlapping stages of pyrolysis: distillation, visbreaking, and coking. Athigher temperatures (250 to 300 degrees, mild cracking of the oil occurs(visbreaking) in which the hydrocarbons lose small side groups and hydrogenatoms to form more-stable, less-branched compounds. At temperatures above about300 degrees C, the residual oil cracks into a volatile fraction and anonvolatile heavy residue of coke, tar, and pitch, which constitutes theprimary fuel for combustion. In the combustion zone, exothermic heterogeneousreactions occur between oxygen in the gas phase and the heavy residue of oildeposited on the rock matrix at lower temperatures.
Summary. The use of additives to improve both steamdrive and cyclic steam injection in field projects has been tested under a variety of conditions. This technique attempts to reduce gravity override and channeling of the steam by foam generation. Another mechanism appears to be "detergent" cleaning near wellbores by surface-active agents. When successful, this technology seems to be economic even at a low oil price. The results, however, have ranged from excellent to negative. In this paper, we attempt to evaluate the field projects published to date. The results of this study show that the use of additives with steam can provide significant benefits over the use of steam alone. Indeed, addition of surfactant to the steam has proved to be both technically and economically successful when the proper products and procedures were used. Caustics have given mixed results but seem to have been effective in at least one cyclic-steam project.
The late 1970's and the 1980's have seen many attempts to improve steam injection by the use of additives. Although other additives have been tried in the field, the most promising technology seems to be injecting a solution of aqueous surfactant either to increase the pressure gradient across the region of interest by generation of foam or to use the detergency properties of the surfactant to reduce the oil/water interfacial tension(IFT) and to modify the relative permeability curves. In our opinion, both mechanisms are present to some degree in most steam-foam projects. Sixteen field tests of steam with additives have been reviewed. They cover a broad spectrum of reservoirs, oils, depths, and pressures. The type of additive used, the surfactant concentration, the presence or absence of noncondensable gases and their nature, and the additive injection mode also varied widely. The effect of these factors on the efficiency of the projects is discussed. Some of the field tests have been performed after extensive laboratory work. The data published on laboratory screening of additives and supporting field data are used to explain some of the results observed in field testing. This paper obviously does not cover all the field projects on steam with additives. Some companies choose not to publish their results. We know that successful projects were not reported. It is possible, however, that a slight bias in favor of success exists in the published literature. In the field, evidence of mobility control has been sought through pressure measurements, tracer testing, injectivity profiling, logging, temperature monitoring, and observation wells. Production data have, of course, been analyzed. The monitoring techniques are also evaluated. Of special interest is the use of these monitoring techniques to identify the recovery mechanisms. Laboratory studies aimed at the determination of surfactant thermal stability and the optimization of a steam/surfactant system are numerous. One of the earliest was probably the study by Gopalakrisnan et al. in which the emphasis was mostly on IFT lowering. This study demonstrated the need to optimize the injection technique to obtain maximum oil recovery in the laboratory. Dilgren et al. tested surfactants at steam-injection conditions and measured the mobility reduction caused by the surfactant compared with water alone. They also emphasized the use of NaCl to improve the foam and noncondensable gases to help stabilize it. The improvement caused by noncondensable gases was demonstrated again by Falls et al. and Janssen-Vanrosmalen et al. Several studies were completed at the Stanford U. Petroleum Research Inst. (SUPRI) on foam flow in sandpacks and surfactant stability at steam-injection conditions. Refs. 8 through 12 also present interesting laboratory studies on various aspects of the steam-foam process. In addition, several of the field tests that will be described followed extensive laboratory work; the main features of the results are described in the papers.
Several trends can be identified from the laboratory studies at this stage. 1. A large number of commercially available surfactants are thermally stable and can be used in steam injection. 2. When mobility control is desired, injection of a noncondensable gas with the surfactant solution increases the mobility reduction of the steam. 3. Introduction of air or oxidants can cause rapid degradation of the surfactant. 4. The addition of brine has a positive effect on some surfactant systems (notably alpha olefin sulfonates), while it has a negative effect on others. 5. Some surfactants will affect only the gas mobility, while others also will modify the interfacial parameters between oil and water and will affect the oil/water relative permeabilities. This point is discussed in detail by Ziritt. In addition, optimum foaming and optimum interfacial properties do not occur simultaneously. A complete discussion of the laboratory work on steam surfactant systems is inappropriate in this paper. It is important, however, to note that most of the successful field projects presented here have benefited from extensive laboratory testing.
Shell Kern River Leases (Mecca and Bishop)
Dilgren et al. presented some preliminary results of a field test in the Mecca lease of the Kern River field. Table 1 shows some of the reservoir parameters. The problem seems to be the classic case of gravity override in a relatively thin, shallow, flat reservoir. The test was performed on four 2.5-acre five-spot patterns. A different surfactant was injected in three wells for testing. Three separate surfactants [Siponate DS-10; Siponate A-168, and Bioterge AS-30( 1618)] were injected continuously at a concentration of 0.5 wt% in the liquid. Salt was added at a concentration of 4 wt% and nitrogen was continuously injected at 3.6 scf/min (about 0.005 mole fraction of the gas phase). The field had been under steamdrive for 10 years before the steam-foam project began. Injection pressure rose from 20 to 40 psi before surfactant injection to 110 to 140 psi with Siponate DS-10, to 155 to 170 psi with Siponate A-168, and to 190 to 205 psi with Bioterge AS-30. Pressure gradients in the drive showed a four- to seven-fold increase. The production of the pilot exhibits an increase of 32,000 bbl over anticipated production by steamdrive alone. It was decided to use Bioterge AS-30 for further work. This surfactant was later replaced by alpha olefin sulfonates. Recently, Patzek and Koinis gave further results of the Mecca lease pilot and added data from the Bishop Fee pilot, also in the Kern River field. The pilot consisted of four contiguous inverted five-spots covering 12 and 14 acres, respectively. While the Mecca pilot was a mature steamdrive (10 years), the Bishop lease had been under steamdrive for only 1 year. Both projects were well-instrumented (eight observation wells in each pilot). Continuous injection of steam, nitrogen, alpha olefin sulfonate, and salt resulted in increased injection pressure.
Copyright 1991 Society of Petroleum Engineers